Our results showed that chest CT imaging is a reliable testing method for COVID-19 evaluation; the higher the blood glucose level is, the more evident the lung inflammation and the more severe the clinical manifestation.

The proportion of patients with diabetes has been growing in the older population [13]. High blood glucose causes an increase of dicarbonyl molecules in the human body, inhibits the antibacterial and immune function of β-defensins and increases the susceptibility of patients, particularly among those with uncontrolled or poorly controlled blood glucose [14]. Previous studies have shown that high blood glucose reduces peripheral blood lymphocytes and increases leukocytes in patients with COVID-19, leading to immune system overactivation and more severe clinical symptoms [5, 15]. Although some of the research indicates that neutrophils and leukocytes help distinguish asymptomatic and moderate COVID-19 [16], this study found that the differences in lymphocytes and leukocytes among patients are not statistically significant.

According to the latest version of the Guidelines on Diagnosis and Treatment of Novel Coronavirus Pneumonia, the diagnostic criterion for COVID-19 is a positive RT-PCR test; however, the detecion takes about one day and the production of test kits is limited. Comparatively, a chest CT scan is an important tool for the rapid screening and initial diagnosis of COVID-19; given its convenience, it can serve as a predictive indicator to evaluate illness severity and clinical prognosis [12, 17, 18] and help in the timely adjustment of patients’ diagnosis and treatment [19]. Most patients show abnormal manifestations of the lungs 2–3 days after contracting COVID-19. In the early stage of the disease, the virus tends to invade peripheral blood vessels and bronchioles, causing interstitial changes in lung tissue, such as interstitial inflammatory oedema and interlobular septal thickening, leading to a rise in pressure of lung parenchyma and exudation of fibrinous and high-protein mucus within bronchiole and thus the formation of GGO and halo sign in the subpleural lung region. Over time, interlobular septal thickening further causes the formation of crazy-paving signs, restricting the absorption of alveolar exudate and causing alveolar consolidation and in serious cases, even diffuse alveolar damage, eventually leading to a white lung and pleural effusion in some patients [18, 20, 21]. A review of relevant literature reveals that compared with patients without diabetes, patients with diabetes show more evident inflammation on chest CT and those with poor blood glucose control exhibit more severe lung damage [3, 5]. Similar to previous studies, the majority of patients in this experiment demonstrated GGO and lung consolidation on their CT appearance, and a few even showed pleural effusion and crazy-paving signs. However, the differences in these changes are statistically insignificant between the two groups, which may be attributed to the small number of cases and good treatment received by most patients during the early onset of the disease.

Since the outbreak of COVID-19, chest CT imaging has played an important role in evaluating infected patients. Pan et al. [22] proposed semiquantitative CTSSs based on the degree of lobar involvement (0–25). We employed the CTSS to evaluate the severity of patients’ images (Fig. 4). Previous research shows that patients with poor blood glucose control have a significantly higher CTSS than those with good blood glucose control [23, 24] because high blood glucose inhibits the immune system and increases the generation of inflammatory factors, causing more severe chest manifestations after the patient contracts viral pneumonia. Statistics show that patients with diabetes have a significantly higher CTSS than those who are not diabetic, and the higher the blood glucose level is on the date of admission, the more severe the lung damage [7, 25, 26]. Our results indicate that blood glucose control is significantly positively correlated to CTSS; patients with poorly controlled blood glucose have higher scores, and lesions commonly involve the lower lobes and peripheral zones of both lungs, further validating that blood glucose affects lung manifestations of the disease. However, blood glucose control and FPG level cannot be treated as independent predictors of clinical severity, and the relationship between blood glucose and COVID-19 requires further exploration. The CTSS can be used to evaluate and predict clinical classification and, in the meantime, offers some accuracy in predicting patient mortality [27], reminding us of the necessity of imaging examination in the diagnosis and treatment of patients with COVID-19. One Iranian study found no significant differences in clinical outcomes and chest CTSSs between patients with diabetes with good and poor blood glucose control [28], which is inconsistent with the findings of this study. This may be attributed to the fact that the former study failed to confirm the diagnosis with test kits and only referred to clinical characteristics and chest imaging manifestation when including patients in the positive group, as well as potential physical differences between Iranian and Chinese patients [29].

Patients’ chest CTSS: Figure A shows mild pneumonia in patients with scattered multiple ground glass opacities in both lungs, the right upper lobe lesion accounting for 13.9% of lobe volume and a CTSS score of 2. Figure B shows a diffuse exudative consolidation scattered in both lungs, with the left upper lobe lesion accounting for 91.4% of lobe volume and a CTSS score of 5. (A) male, 48 years old, a history of diabetes for 3 years, regular medicine with well-controlled blood glucose. (B) female, 44 years old, with history of diabetes for 5 years, irregular medication and poor blood glucose control

Our study also had some limitations. First, the retrospective design with a relatively small sample size may not be adequate to get reliable results. Moreover, COVID-19, as a newly discovered disease, still has many mysteries that remain in dispute, including the assessment of severity. Further large-scale prospective design studies should be conducted to explore the characteristics of COVID-19 and its relationship with diabetes. Moreover, we only included patients with diabetes; the results in patients without diabetes or with other diseases should be explored and the number of patients with COVID-19 infection was not clear in the research hospital.

In summary, chest CT imaging is a reliable testing method for COVID-19 evaluation; the higher the blood glucose level is, the more evident the lung inflammation and the more severe the clinical manifestation. This indicates that clinical patients should first receive a CT scan upon admission to evaluate their initial condition, and imaging appearances should be analysed thoroughly. Targeted treatment should be administered, and publicity and education activities should be implemented aiming to popularise among patients with diabetes the necessity of controlling their blood glucose and encouraging them to make an active effort to keep healthy and follow appropriate diets, thus improving their quality of life.